Multi‐omics molecular phenotyping reveals the potential mechanisms of chemotherapy response and resistance in small cell lung cancer

Dear Editor, Small cell lung cancer (SCLC) is a low-survival malignant lung cancer with mainly extensive stage (ES).1, 2 A major challenge in treating SCLC is chemotherapy resistance.3 However, studies on disease evolution and molecular mechanisms of resistance during chemotherapy are insufficient. Here, we conducted a multicentre, observational study to profile the multi-omics characteristics of tumour tissue, circulating tumour cell (CTC) and circulating tumour DNA (ctDNA) in Chinese ES-SCLC patients. This study enrolled 54 patients, including naïve cohort and relapsed cohort (Figure S1 and Table S1). Except for one patient who had distant metastasis in relapsed cohort, all the patients had ES-SCLC. According to the different stratification parameters, patients were divided by two manners. One manner was chemo-resistant vs chemo-sensitive, according to whether the time from the end of first-line therapy to disease progression exceeded 90 days (chemotherapy-free interval); other manner was responders vs non-responders, according to Response Evaluation Criteria in Solid Tumours (version 1.1), and the patient whose lesions shrank over 30% was defined as responder. The median overall survival (OS) and progression-free survival (PFS) for all patients were 9.6 m (95% confidence interval [CI]: 7.5‒12.2 m) and 4.5 m (95% CI: 3.4‒5.7 m), respectively. No clinical parameters had a significant effect on prognosis (Table S2). Chemo-sensitive/response patients had longer PFS (Figure S2), which suggested that different biological contexts may exist. The detection of ctDNA mutations was highly consistent with tumour results and the tumour mutation burden (TMB) was highly correlated (Figure S3A‒D and Tables S3 and S4), which indicated that ctDNA mutations could be used to monitor mutational changes during treatment with high confidence. As expected, TP53 and RB1 mutations were detected in most patients' baseline ctDNA (Figure S3A). Some frequently deleted genomic regions in tumours and more in CTCs were found (Figure S3E,F), which may indicate the evolution of genomic heterogeneity among diverse clones and the initial development of drug resistance. The tumour showed a high proportion of C > A transitions (Figure S3G). In both non-responders and chemo-resistant in baseline ctDNA, only the KDR gene (vascular endothelial growth factor receptor [VEGFR]) had a significantly higher mutation frequency (Figure 1A,C). The baseline ctDNA of chemo-resistant showed more significant deletion frequency, but only SORCS1 had a significant deletion frequency in baseline tumours (Figure S4). The TMB of baseline ctDNA in non-responders and microsatellite instability (MSI) score of baseline tumours in chemo-resistant were significantly higher (Figures 1B,D and S7B). However, other genomic indexes in tumour had no significant differences (Figures S5‒S7). Three pathways were highly enriched and one pathway was lower in non-responders (Figure S8). Tumour samples clustered into high and low levels of immune infiltrate by RNA-sequencing (Figure 2A,B). Although immune infiltration had no significant difference (Figure 2C,D), some immune populations in non-responders/chemo-resistant were significantly higher (Figures 2E, S9 and S10). The KRAS signalling pathways were enriched in non-responders in tumour (Figures 2F and S11), which were reported to affect the presence and suppressive function of tumouricidal cells.4 Conversely, several pathways related to proliferation and immunity were significantly up-regulated in responders/chemo-sensitive in both baseline tumours and CTCs (Figures 2F‒I, S11 and S12), suggesting that a more vital ability of differentiation and immunogenicity may occur in chemotherapy-sensitive tumour. The cell death pathway related to pyroptosis was different in baseline CTCs, and there was a significantly higher score of alkaliptosis in relapse nodes (Figure 2J,K). After chemotherapy, the CTCs in chemo-sensitive were significantly reduced at C3D1 while fewer changes were observed in chemo-resistant (Figure S13). The changes in the CTC counts and molecular tumour burden index (mTBI)5 were nearly consistent during conventional follow-up. The responders mostly tended to show a decreasing trend, while chemo-resistant showed a more frequent increasing trend (Figure S14). Phylogenetic relation trees showed a sustained high cancer cell fraction of major clones in chemo-resistant in both baseline and relapsed samples, but chemo-sensitive was characterised by the weakening of major clones in baseline samples (Figure S15A,B). Although the average number of mutations of trunk private clones was significantly higher in non-responders, the fraction of functional genes was lower (Figure S15C,D). The genomic landscape of 21 paired baseline and relapsed ctDNA showed no significant differences in TMB and mTBI (Figure S16). The KDR gene was still one of the top 10 frequently mutated genes (Figure 3A). The platinum drug resistance pathway was significantly enriched in baseline subclonal mutations and relapsed clonal mutations (Figure 3B), which indicated that tumour with drug-resistant mutations expanded from subclone to clone. Meanwhile, the tyrosine kinase inhibitor resistance and immune-related pathways were enriched in relapsed clonal mutations and subclonal mutations, respectively. Finally, we summarised the correlation between mutation/pathway/immunity and pathological response or chemotherapy sensitivity (Figure 3C), which may provide a comprehensive concept of treatment response and resistance mechanisms in SCLC. Consistent global copy number variation (CNV) results from cell lines and patient 1022 were observed, and some significant CNV changes were found between patients with or without durable clinical benefit (Figure S17). Patients with KDR mutation tended to have higher KDR expression levels and poor prognosis (Figure S18A,B). By using other datasets, we found that KDR was significantly highly expressed in SCLC-I and patients with low expression of the KDR were enriched in SCLC-A in the IMpower133 cohort6 (Figure S18C‒F). OS was significantly shorter in patients with high expression of KDR and VEGF pathways (Figure 4A,B). Moreover, the tumour with a high expression of KDR tended to be 'hot' (Figures 4C,D and S19). These findings were consistent with previous pathway enrichment results, suggesting that patients with KDR mutation and/or high expression of KDR may resist chemotherapy but benefit from immunotherapy, anti-folates and AURK inhibitors.7 Several chemotherapy agents contained higher IC50 in the high KDR expression group (Figure 4E), which may reveal a resistant trend. In conclusion, chemo-sensitive/response patients showed beneficial survival, and we found the potential mechanism was that KDR mutation, PI3K amplification, VEGF and KRAS pathways activation contribute to the development of chemotherapy resistance. These results provide critical information for the clinical decision of VEGF signalling pathway inhibitors combined with chemotherapy and imply that targeted therapies may benefit some patients who are resistant to chemotherapy. Besides, the difference in tumour microenvironment and several immune pathways enrichment between chemo-sensitive and chemo-resistant was consistent with the concept that tumours can take control of environment to reset the body homeostasis.8 However, we still lack sufficient evidence to determine the most appropriate therapies for recurring patients. Future studies are warranted on larger cohorts of patients in a real-world cohort to explore. Conceptualisation, supervision, funding acquisition and writing—review and editing: Ying Cheng. Resources, data curation, software, formal analysis, methodology, writing—original draft and writing—review and editing: Xuan Gao and Zelong Xu. Formal analysis, methodology and writing—review and editing: Bingfa Yan. Conceptualisation, resources and writing—review and editing: Jie Hu. Resources, data curation and writing—review and editing: Ying Liu, Jing Zhu, Ying Wang, Junfeng Wang, Changliang Yang, Hongxia Cui, Yanrong Wang, Guang Yang, Jie Hao, Peidong Li, Liang Zhang, Zili Li, Hongyu Wang, Yanli Sun, Shubo Zuo and Tianying Du. Software, formal analysis and writing—review and editing: Zhentian Liu, Xuefeng Xia and Xin Yi. Resources and writing—review and editing: Ying Xin, Ke Zheng, Yawen Yang and Kai Niu. Formal analysis and writing—review and editing: Jinhua Xu, Gan Zhang, Fei Chen and Ning Ding. We are greatly thankful for the funding for this study provided by the Development and Reform Commission of Jilin Province (2021C043‑1) and the Science and Technology Planning Project of Jilin Province (YDZJ202202CXJD009). We greatly appreciate the patients and investigators who participated in this study for providing the data. Xuan Gao, Bingfa Yan, Zelong Xu, Zhentian Liu, Xuefeng Xia and Xin Yi are employees of Beijing GenePlus Technology Co., Ltd. The remaining authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as potential conflicts of interest. All patients provided written informed consent to conduct research in this study, and ethical approvals were obtained from the two hospitals (NOPRODLUC0001). The study received approval to conduct genomic research from the China Human Genetic Resources Administration Office (HGRAO, 2016-161). Please note: The publisher is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.